Skip to content

Provision an EKS Cluster - How to

On this page you will learn how to setup an edge system with an Amazon Kubernetes Service (EKS) cluster using the cluster installation tools in the Kelvin UI.

If you want Kelvin to be responsible to also install the Kubernetes cluster directly and handle all the management of the cluster, you can choose the K3S option.

Amazon Kubernetes Service (EKS)

Amazon Elastic Kubernetes Service (Amazon EKS) is a fully-managed service that simplifies the deployment and management of Kubernetes clusters. It provides a highly available and scalable platform to deploy containerized Kelvin SmartApps™ using Kubernetes, without the need to manage the underlying infrastructure.

Requirements & Limitations

Cloud Kubernetes implementations vary across providers, and even within the same provider, there can be different types of stacks based on the selected services.

So it is not possible to list all the requirements, limitations and fixed design options available and will heavily depend on your project and budgets.

When setting up your cluster on EKS here are some links to the EKS documentation that can help you decide what type of setup to implement.

Installation

This step will install all the Kelvin services to your EKS cluster.

Setup EKS Cluster

To start you need to setup your own EKS cluster following Amazon's instructions. There are four basic steps to installing and setting up EKS clusters;

  1. Create an EKS Cluster: Use the AWS Management Console, AWS CLI, or an AWS SDK to create an EKS cluster.
  2. Configure kubectl: Set up kubectl, the Kubernetes command-line tool, with credentials to access your EKS cluster.
  3. Create Worker Nodes: Launch and configure worker nodes that will join your cluster. This can be done through AWS Management Console or using AWS CloudFormation templates.
  4. Connect Worker Nodes to Cluster: After creating the worker nodes, connect them to your EKS cluster.

Kuberenetes documentation website has a getting started guide for achieving all these steps; https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/

Install Cluster

To start, go to the Orchestration page and click on the Register Cluster button;

Then in the popup options, select Kubernetes and type in a display name and optionally a name ID for this new EKS cluster.
Display Name This can be any characters and spaces that gives your cluster a memorable name for reference
Name ID A unique lower-case alphanumeric name which uniquely identifies this cluster. This will be automatically filled in when you type the Display Name. Normally you do not need to change unless the Name ID clashes with another cluster's Name ID.

The Name ID is the unique identifier name for the cluster. This must contain only lowercase alphanumeric characters. The ., _ and - characters are also allowed to separate words instead of a space BUT can not be at the beginning or end of the name.

You can then download the manifest YAML file and then run the CLI command in your AKS environment.

Make sure the manifest YAML file is locally accessible to the kubectl command.

The actual CLI command is;

kubectl apply --server-side -f kelvin.yaml

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.

curl -X "POST" \
"https://<url.kelvin.ai>/api/v4/orchestration/clusters/create" \
-H "Authorization: Bearer <Your Current Token>" \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-d '{
    "name": "doc-demo-cluster-eks",
    "title": "Doc Demo Cluster EKS",
    "type": "kubernetes"
}'

The response will look something like this;

{
   "name":"doc-demo-cluster-eks",
   "title":"Doc Demo Cluster EKS",
   "type":"kubernetes",
   "ready":false,
   "status":"pending_provision",
   "last_seen":null,
   "sync_scrape_interval":120,
   "manifests_scrape_interval":130,
   "manifests_scrape_enabled":false,
   "version":{
      "k8s_version":"",
      "kelvin_version":""
   },
   "created":"2024-05-23T14:20:59.62627Z",
   "updated":"2024-05-23T14:20:59.62627Z",
   "service_account_token":"bm9kZS1jbGllbnQtZG9jLWRlbW8tY2x1c3Rlci1la3M6VE9VMWZsSmEwRUtNODZSSWM0RnI1ZFBXRDNYWjlnaUg=",
   "provision_script":"kubectl apply --server-side -f kelvin.yaml",
   "join_script":"",
   "telemetry_scrape_interval":110,
   "telemetry_enabled":false,
   "telemetry_buffer_size":10,
   "forward_logs_enabled":false,
   "forward_logs_buffer_size":10,
   "upgrade_status":{
      "state":"idle",
      "message":""
   },
   "upgrade_pre_download":false,
   "upgrade_instantly_apply":true
}

You can then get the manifest YAML file that is mentioned in the provision script with the API request;

curl -X "GET" \
"https://<url.kelvin.ai>/api/v4/orchestration/clusters/doc-demo-cluster-eks/provision/get" \
-H "Authorization: Bearer <Your Current Token>" \
-H "Accept: application/json"

which will give a YAML response starting like this;

kind: Namespace
apiVersion: v1
metadata:
  name: kelvin-admin
---
apiVersion: v1
kind: Secret
type: kubernetes.io/tls
metadata:
  name: kelvin-certificate
  namespace: kelvin-admin
  labels:
    app.kubernetes.io/part-of: kelvin
data:
  ca.crt:

  ...

Copy and paste this into a file called kelvin.yaml and then run the CLI command given in the key provision_script in your EKS environment.

Make sure the manifest YAML file is locally accessible to the kubectl command.

The actual CLI command is;

kubectl apply --server-side -f kelvin.yaml

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.

from kelvin.api.client import Client

# Login
client = Client(config={"url": "https://<url.kelvin.ai>", "username": "<your_username>"})
client.login(password="<your_password>")

# Set or Update the Default Cluster in an Instance
response = client.orchestration.create_orchestration_clusters(data={
    "name": "doc-demo-cluster-eks",
    "title": "Doc Demo Cluster EKS",
    "type": "kubernetes"
    })

print(response)

You can save the manifest YAML file that is mentioned in the provision script with;

kelvin_yaml = client.orchestration.get_orchestration_clusters_provision(cluster_name=response.name)
with open("kelvin.yaml", 'w') as file:
    file.write(str(kelvin_yaml))

Make sure the manifest YAML file is locally accessible to the kubectl command.

And then run the CLI command given in the key provision_script in your EKS environment. The actual CLI command is;

kubectl apply --server-side -f kelvin.yaml

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.